Thiourea inhibitors of herpes viruses. Part 2: N-Benzyl-N′- arylthiourea inhibitors of CMV

Article abstract of DOI:10.1016/j.bmcl.2004.04.093

A series of highly potent thiourea inhibitors of cytomegalovirus (CMV) with improved stability properties was prepared and evaluated. Compound 29 inhibited the virus in cultured HFF cells with IC₅₀ of 0.2nM.

Full text of DOI:10.1016/j.bmcl.2004.04.093

Bioorganic & Medicinal Chemistry Letters 14 (2004) 3401–3406
Thiourea inhibitors of herpes viruses. Part 2: N-Benzyl-N0-
arylthiourea inhibitors of CMV
Jonathan D. Bloom,* Russell G. Dushin, Kevin J. Curran, Fran Donahue,
Emily B. Norton, Eugene Terefenko, Thomas R. Jones, Adma A. Ross,
Boris Feld, Stanley A. Lang and Martin J. DiGrandi
Chemical and Screening Sciences and Infectious Disease Research, Wyeth Research, 401 N. Middletown Rd,
Pearl River, NY 10965, USA
Received 29 January 2004; accepted 27 April 2004
Available online
Abstract—Aseries of highly potent thiourea inhibitors of cytomegalovirus (CMV) with improved stability properties was prepared
and evaluated. Compound 29 inhibited the virus in cultured HFF cells with IC50 of 0.2 nM.
Ó 2004 Published by Elsevier Ltd.
We recently reported the preparation and evaluation of
a series of novel bis(aryl)thiourea inhibitors of cyto-
megalovirus (CMV).1 The compounds, as exemplified
by structures 1a,b (Fig. 1), were found to be highly
potent and specific inhibitors of CMV in cultured
human fibroblasts. Unfortunately, as pre-development
studies progressed we noted an unexpectedly facile
decomposition of the thiourea moiety, especially in
acidic media under forcing conditions. Since CMV re-
mains an important infection in certain populations,2–4
we embarked on an effort to prepare more suitable
compounds that would retain anti-CMV activity while
exhibiting greater hydrolytic stability.
aniline5 nitrogen and subsequent elimination of the
corresponding aniline 2 and isothiocyanate 3. We fur-
ther assumed that this degradation was facilitated by the
electronegative nature of the trifluoromethyl groups,
augmenting the leaving group ability of the aniline. The
placement of an appropriate spacer between the aniline
nitrogen and the phenyl ring would be expected to
modulate this effect and therefore improve the stability
characteristics of the molecule. Aseries of N-benzyl and
N-a-alkylbenzyl analogs of type 4 was prepared.
Thioureas 4 were easily prepared as shown in Scheme 1
by reaction of amines 5 with isothiocyanates 61 in hot
acetonitrile.
We reasoned that the decomposition (Fig. 2) likely re-
sulted from protonation of the 3,5-(bis)trifluoromethyl-
The preparation of amines 5 is shown in Scheme 2. In
the case where W ¼ CH2 the required 3,5-bis(tri-
fluoromethyl)benzylamine was commercially available.
a-Methyl- benzylamines 5, (W ¼ CH(CH3)) was prepared
by conversion of commercially available 3,5-bis(trifluo-
romethyl)phenyl methyl ketone 7 to the O-methyl oxime
9 and subsequent reduction using the conditions reported
by Itsuno et al.7 Alternatively, using the conditions
described by Leukhardt,8 7 could be converted to the
2-(formamido)ethylbenzene 8, which yielded 5 upon basic
hydrolysis. Enantiomerically pure S-(5a) or R-a-methyl-
3,5-(bis)trifluorobenzylamine (5b) could be prepared by
CF3
X
S
N
H
N
S
C
O
F3C
N
H
N
H
1a (X = CH)
1b (X = N)
Figure 1. Discontinued CMV inhibitors.
fractional crystallization of the
respectively. The absolute configuration of 5a and 5b was
L- or D-tartrate salt,
* Corresponding author. Fax: +1-845-602-5561; e-mail: bloomj@
wyeth.com
determined according to the Mosher amide method.9
0960-894X/$ - see front matter Ó 2004 Published by Elsevier Ltd.
doi:10.1016/j.bmcl.2004.04.093

3402
J. D. Bloom et al. / Bioorg. Med. Chem. Lett. 14 (2004) 3401–3406
X
N
S
H
N
O
CF3
S
CF3
H
N
N
X
S
C
N
F3C
H
S
F3C
N
H
H
C
H
HN
O
N
H
1
CF3
X
S
N
H
N
O
S
C
N
F3C
NH2
3
2
H
H
N
N
Ar
Ar
S
C
S
C
O
O
H
H
W
HN
N
N
N
H
H
H
F3C
W
F3C
CF3
CF3
4
X
CF3
H
N
Ar
NH2
O
F3C
W
S
C N
5
6
Figure 2. Proposed mechanism of thiourea decomposition.
X
N
S
CF3
SCN
i
5
O
N C
+
H
N
H
N
H
N
H
O
N
X
W
F
3C
S
S
4
6a (X = CH)
6b (X = N)
Scheme 1. Reagents and conditions: (i) hot acetonitrile.
CF3
CF3
CF3
i
CF3
ii
CH3
NHCHO
iii
NH2
CH3
F3C
F3C
8
CH3
NH2
CH3
5a
F3C
F3C
O
7
v
CF3
CF3
vi
5
iv
CH3
NH2
CH3
F3C
F3C
NOCH3
5b
9
Scheme 2. Reagents and conditions: (i) ammonium formate, formic acid, formamide; (ii) NaOH, H2O; (iii)
(iv) -tartaric acid, EtOH crystallization; (v) MeONH2ÆHCl, pyr. EtOH; (vi) NaBH4, ZrCl4.
L-tartaric acid, EtOH crystallization;
D

J. D. Bloom et al. / Bioorg. Med. Chem. Lett. 14 (2004) 3401–3406
3403
respectively. Reduction of the nitro group using iron and
ammonium chloride afforded the aminopyridines 17a,b,
which were converted to the corresponding isothiocya-
nates 18a,b upon treatment with 1,10-thiocarbonyldiimi-
dazole. Reaction with amines 5a,b provided the thioureas
19a–c as described above.
CF3
CF3
i
R
F3C
F3C
CHO
OH
11
ii
10
The stabilizing influence of an alkyl spacer group is
illustrated (Table 1) by comparing the rate of degrada-
tion of compounds 1 and 28 under forcing conditions.
CF3
CF3
iii
R
R
F3C
F3C
Table 1. Comparative stability of thioureas 1a and 28
NH2
O
13
12
CF
3
O
C
Scheme 3. Reagents and conditions: (i) RMgCl/Et2O; (ii) Jones oxi-
dation; (iii) see Scheme 2.
N
H
N
H
N
H
N
F C
3
W
S
S
More highly substituted ketones 12 were prepared as
shown in Scheme 3. This was accomplished by addition
of the appropriate Grignard reagent to 3,5-bis(trifluoro-
methyl)benzaldehyde 10 to give alcohols 11. Jones oxi-
dation gave the elaborated ketones 12, which were
converted into the a-substituted benzylamines 13 as
described in Scheme 2. The O-methyl oxime/reduction
sequence7 also worked well for preparation of interme-
diates 12.
1a (W = bond)
28 (W = S-CH(Me))
Compound, conditions
% Remaining
1
a
28
Water @ 37 °C
1 Day
7 Days
1 Day
7 Days
1 Day
7 Days
1 Day
7 Days
1 Day
7 Days
89
44
11
0
100
99
Water @ 60 °C
ND
75
0.1 N HCl @ 37 °C
1.0 N HCl @ 37 °C
0.1 N NaOH @ 37 °C
71
22
52
14
98
97
100
96
Compounds in which the central phenylenediamine
functionality was replaced by the corresponding diamino-
pyridine group were prepared according to Scheme 4.
2-Amino-4-nitropyridine 14 was heated in THF with
excess acid chlorides 15a,b to yield the amides 16a,b,
88
34
100
99
X
S
COCl
N
O N
2
O N
2
N
i
X
N
N
H
O
N
NH
2
S
14
16a (X = CH)
16b (X = N)
15a (X = CH)
15b (X = N)
X
S
X S
N
N
H N
2
SCN
iii
ii
N
N
H
O
N
N
H
O
17a (X = CH)
17b (X = N)
18a (X = CH)
18b (X = N)
CF
3
iv
O
N
H
H
N
H
N
N
C
X
F C
3
N
S
CH
S
3
19a (X = CH, R-enantiomer)
19b (X = CH, S-enantiomer)
19b (X = N, S-enantiomer)
Scheme 4. Reagents and conditions: (i) refluxing THF; (ii) Fe/NH4Cl; (iii) 1,10-thiocarbonyldiimidazole; (iv) 5a, hot acetonitrile.

3404
J. D. Bloom et al. / Bioorg. Med. Chem. Lett. 14 (2004) 3401–3406
CF3
H
N
N
CF3
H
N
N
S
H
N
N
S
EtOH
H+
F C
3
S
S
S
O
C
N
H
O
HN
O
SCN
F3C
NH2
N
H
EtO
3a
3
2
1
X
CF3
H
N
N
S
O
NCS
F C
3
H2N
3b
3c
Figure 3. Ethanol trapping experiment.
As expected, significant increases in stability were noted
in the a-methylbenzyl analog 28 under neutral, mildly
acidic, and strongly acidic conditions when compared to
1.6 These data support the assumption that the mecha-
nism of the rapid decomposition of 1 is acid catalyzed
and facilitated by the trifluoromethyl groups. The
observation that 0.1 N NaOH retarded the decomposi-
tion of 1 when compared to water alone is further sup-
port for the proposed acid-catalyzed mechanism.
One possible weakness of the proposed mechanism is
that it requires that the weakly basic thiourea nitrogens
be protonated. Although protonation would certainly
promote the forward reaction, it is also possible that the
rate-determining step is, in fact, the breakdown of the
tetrahedral intermediate resulting from the direct addi-
tion of water to the thiocarbonyl group. This break-
down could then be inhibited by ionization of the (bis)-
trifluoromethyl aniline NH bond under basic conditions
and subsequent decreased leaving group capability. This
alternative mechanism would also explain the base sta-
bility of the compounds.
Further supporting evidence for the above mechanism
was derived from an ethanol trapping experiment (Fig. 3)
in which 1 was exposed to anhydrous acidic ethanol.
After 1 day, the reaction products were identified by LC/
MS. Aniline 2 and isothiocyanate 3 were identified, in
addition to 3a, the ethanol adduct of 3. Neither aniline
3b nor isothiocyanate 3c were detected. These data
support the above mechanism and suggest that while
both thiourea nitrogens are probably reversibly proton-
ated, the rate determining step is the elimination of the
amino group, which is faster in the case of 2 group than
for 5, presumably due to the difference in leaving group
ability.
All compounds cited herein (Tables 2–4) were tested10
against the herpes viruses CMV, HSV (herpes simplex),
and VZV (varicella zoster). Additionally, screening
against non-herpes viruses such as RSV (respiratory
syncytial virus) and analysis of MTS host cell toxicity11
were performed to distinguish between specific anti-viral
activity and nonspecific host cell toxicity.
In addition to demonstrating significant improvement in
hydrolytic stability compared to the earlier analogs, the
Table 2. Thiadiazole analog SAR
CF
3
O
C
N
H
N
H
N
H
N
N
S
F C
3
W
S
Compound
W
CMVa
HSVa
VZVa
RSVa
MTSa
1b
20
21
22
23
24
25
26
None
CH2
0.037
0.002
0.003
0.003
0.001
0.001
0.002
0.004
>5
>5
>5
>5
5
>5
>5
5
>5
>5
>5
>5
>5
>5
>5
ND
ND
rac-CH(CH3)
S-CH(CH3)
CH(Et)
CH(n-Pr)
CH(Ph)
CMe2
>3.7
>3.7
>3.7
>3.7
>3.7
>3.7
>7.5
>7.5
>7.5
>7.5
>7.5
>7.5
>5
>5
>5
ND ¼ Not determined.
a IC50, lM.

J. D. Bloom et al. / Bioorg. Med. Chem. Lett. 14 (2004) 3401–3406
3405
Table 3. Thiazole analog SAR
CF
3
O
C
N
H
N
H
N
H
N
F C
3
W
S
S
Compound
W
CMVa
HSVa
VZVa
RSVa
MTSa
1a
27
28
29
None
0.016
>5
>5
>5
4.1
>5
>5
2.0
>5
>5
ND
R-CH(CH3)
S-CH(CH3)
CH(Et)
0.0015
0.0006
0.0002
>3.7
>3.7
>3.7
>7.5
>7.5
>7.5
ND ¼ Not determined.
a IC50, lM.
Table 4. SAR of pyridyl analogs
Compound
W
X
N
CMVa
HSVa
VZVa
RSVa
>5
1.0
MTSa
30
CH2
0.0044
0.0028
0.0020
0.0030
>5
>5
>5
4.1
>3.7
>3.7
>3.7
>3.7
>7.5
>7.5
>7.5
8
19a
19b
19c
R-CH(CH3)
S-CH(CH3)
S-CH(CH3)
CH
CH
N
>5
>5
a IC50, lM.
spacer-containing thioureas were also significantly more
potent than the earlier compounds, regardless of the
linker chosen (Table 2).
CF
3
O
C
N
H
N
H
N
H
N
X
F C
3
W
Compound 1b was previously found1 to be a highly
potent inhibitor against CMV in cells with IC50 of
37 nM. Inclusion of a methylene spacer (20) improved
the activity 15-fold to 0.002 lM (2 nM). Addition of
single branched alkyl substituents had little effect on
potency. The racemic methyl analog 21 had the same
activity as 20 and the S-enantiomer 22. Elaboration of
the methyl group resulted in a small increase in potency,
with the ethyl (23) and n-propyl (24) analogs being
about twofold more active than 15. The gem-dimethyl
(26) and phenyl (25) analogs had similar activity to 20.
Noting that the minimum MTS and other virus IC50
values for this series is at least 3.7 lM, one can estimate
the selectivity index (SI) to be at least 7500 for the
compounds in Table 2.
N
S
S
Figure 4. Structures of pyridine analogs.
absence of a stereochemical preference at the benzyl
center was once again apparent (19a vs 19b).
This study demonstrated that it was possible to address
the stability issues of a previous series that prevented
development of the compounds as a potential CMV
therapy. The cause of the instability of the original
compounds was uncovered and addressed mechanisti-
cally by addition of spacer groups to stabilize the thio-
urea group. Furthermore, these spacer groups afforded
greater potency to the molecules. Single digit and sub-
nanomolar inhibitors resulted without any appreciable
MTS toxicity or loss of specificity. The most potent
compound, 29, was a 0.2 nM inhibitor of CMV with a
selectivity index of at least 150,000. Unfortunately, all of
the compounds in the series exhibited very poor bio-
availability, which precluded the selection of any
development candidates (Fig. 5).
Asimilar SAR pattern was found in analogs of the
thiazole series (Table 3). As previously reported,1 the
thiazole group affords about fivefold more potency
when compared to the corresponding 1,2,3-thiadiazole
analogs (1a vs 1b, 22 vs 28).
There was little difference between the R-(27) and S-(28)
enantiomers, as shown in Table 3. The branched ethyl
analog 29 was the most potent compound in the series
(0.2 nM).
CF
3
O
C
N
H
N
H
N
H
N
Several pyridine analogs were prepared (Fig. 4) to
improve solubility characteristics of the molecules.
These compounds were about twofold less active than
the comparable phenyl analogs, but still retained sig-
nificant activity and selectivity indices. (Table 4). The
F C
3
S
S
H C
3
Figure 5. Compound 29, a 0.2 nM inhibitor of CMV.

3406
J. D. Bloom et al. / Bioorg. Med. Chem. Lett. 14 (2004) 3401–3406
found to be superior. Other substituents are not included
in this manuscript.
References and notes
6. It should be noted that the a-methyl group plays no
obvious part in this stabilization, as corresponding ana-
logs with a simple methylene spacer show similar stability
characteristics. The selection of a-methyl compounds was
made based on PK considerations.
7. Itsuno, S.; Sakurai, Y.; Ito, K. Synthesis 1988, 995–996.
8. For a review of the Leuckart reaction, see: Moore, M. L.
Org. React. 1949, 5, 301–330.
1. Bloom, J. D.; DiGrandi, M. J.; Dushin, R. G.; Curran,
K. J.; Ross, A. A.; Norton, E. B.; Terefenko, E.; Jones,
T. R.; Feld, B.; Lang, S. A. Bioorg. Med. Chem. Lett.
2003, 13, 2929–2932.
2. Britt, W. J.; Alford, C. A. In Fields Virology; Fields, B. N.,
Knipe, D. M., Howley, P. M., Eds.; Lippincott-Raven:
Philadelphia, 1996; 3rd ed.; Vol. 2, pp 2493–2512.
3. Palella, F. J.; Delaney, K. M.; Moorman, A. C.; Loveless,
M. O.; Fuhrer, J.; Satten, G. A.; Aschman, D. J.;
Holmberg, S. D. New Engl. J. Med. 1998, 338, 853–860.
4. (a) Holland, G. N.; Buhles, W. C., Jr.; Mastre, B. Arch.
Ophthalmol. 1989, 107, 1759–1766; (b) Spector, S. A.;
Weingeist, T.; Pollard, R. B. J. Infect. Dis. 1993, 168, 557–
563.
9. Dale, J. A.; Dull, D. L.; Mosher, H. S. J. Org. Chem. 1969,
34, 2543–2549.
10. To calculate IC50s the compounds were serially diluted
and tested in duplicate. The sigmoidal hill slope (0–100)
model was used for curve fitting and IC50 determination.
Each compound was tested at least in two independent
runs on different days.
5. Although several other substituents (electron withdraw-
ing) on the left-hand phenyl ring are consistent with
excellent CMV activity, the 3,5-bis(trifluoro) group was
11. Barltrop, J. A.; Owen, T. C.; Cory, A. H.; Cory, J. G.
Bioorg. Med. Chem. Lett. 1991, 1, 611–614.